OpenAI announced today that, at the request of the Trump administration, it will delay the public release of its latest frontier AI models — the GPT-5.6 series, which includes Sol, Terra, and Luna. While these models have the potential to assist cybersecurity defenders, OpenAI has decided to hold them back from general availability due to national security concerns. In a press release, the company stated it will first provide a limited preview to “a small group of trusted partners,” a list that has been shared with the government, before rolling out broader access.
OpenAI emphasized that it “believes in broad access” and plans to fully release GPT-5.6 Sol, Terra, and Luna in the coming weeks. As part of ongoing coordination with the U.S. government, OpenAI previewed the models’ plans and capabilities to officials ahead of today’s announcement. The decision comes as a U.S. House committee pushes forward with AI safety legislation, reflecting the growing tension between advanced AI capabilities and security controls in the tech industry.
